The use of antidepressants to treat depressive phases or components of bipolar disorder can neither be condemned nor endorsed without carefully evaluating individual clinical cases and circumstances… No simple guidelines can be provided.1
Few topics in psychopharmacology evoke such vigorous debate as the use of antidepressants in patients with bipolar disorder. The passion with which some authors disparage antidepressant use is matched perhaps only by the astonishing paucity of research on their safety and efficacy for this condition. That trend shows little end in sight as many clinicians and other stakeholders accept as a foregone conclusion the proposition that antidepressants for bipolar disorder are broadly ineffective, hazardous to mood stability, or both. The International Society for Bipolar Disorders (ISBD) Task Force on Antidepressant Use in Bipolar Disorder1 acknowledges that “the available evidence concerning both the value and the risks of antidepressant treatment in bipolar disorder is remarkably limited, and much of it is methodologically weak. Therefore, it is not currently possible to make firm clinical recommendations that are soundly evidence based.” Yet, research on antidepressant safety and efficacy has nearly ground to a halt, as if the existing database is less scant and more definitively negative than the ISBD Task Force thinks.
Risk of Antidepressant-Associated (Hypo)mania
When asked what particular “issues” surround the use of antidepressants in bipolar disorder, many clinicians cite concerns about mania induction or cycle acceleration. However, the literature suggests that distinct affective polarity switches unequivocally triggered by antidepressants are rare. In the largest meta-analysis of this issue to date (comprised of 51 studies with 10,098 participants), the pooled occurrence rate of treatment-emergent mania/hypomania was 18.8%, with a higher incidence among retrospective (30.9%) than prospective (14.4%) or randomized (11.8%) studies.2 But, such risk ratios do not address event risk relative to the spontaneous base rate of post-depressive manias/hypomanias in matched bipolar patients not exposed to antidepressants. An earlier meta-analysis3 comparing prospectively observed switch rates with antidepressants (event rate = 7.7%) versus placebo (event rate = 7.2%) revealed a number-needed-to-harm of 2,003—indicating that for most bipolar patients, antidepressant-associated polarity switches are uncommon.
In retrospective or cross-sectional studies of mania during antidepressant therapy, causal directions are hard to infer if treatment is nonrandomized, and if confounding factors that predispose some patients to polarity switches are not accounted for. The National Institute of Mental Health's Systematic Treatment Enhancement Program for Bipolar Disorder acute bipolar depression randomized trial,4 which found no differences either in efficacy (∼25%) or switch rates (∼10%) during mood stabilizer therapy alone or with an antidepressant, observed that much of what clinicians infer as iatrogenic polarity switches may simply be the natural course of illness. Additional methodological issues to consider when interpreting this literature include the following:
- Suspected “switch” events are often identified retrospectively without formal documentation or operational criteria, potentially missing both false-positive and negative events, misattributing subthreshold symptoms as true episodes, and potentially conflating hypomania with other phenomena such as anxious distress or simple agitation or edginess, akathisia, or substance intoxication or withdrawal states.
- The timeframe for judging likely causal associations between antidepressant exposure and polarity switch is often ignored. The ISBD has proposed counting polarity switches within 8 weeks of antidepressant initiation or dosage increases as “definite,” and those occurring by 12 weeks as “likely,” and those by 16 weeks as “possible.”5 Beyond that timeframe, plausible causality becomes indistinguishable from the natural course of illness.
- It is often assumed that concomitant use of an antimanic mood stabilizer or antipsychotic medication is a necessary prerequisite to antidepressant treatment in bipolar disorder, mainly for presumed protection against induction of mania or hypomania. In fact, no randomized controlled trial (RCT) has ever demonstrated this to be true. A meta-analysis of 35 trials found no evidence that mood stabilizer cotherapy reduces the likelihood of switch events during antidepressant therapy (15.9% risk during antidepressant plus mood stabilizer, 13.8% risk during antidepressant without mood stabilizer).6
Psychopharmacologists often cite lack of efficacy, more than the potential to induce mania, as the greater concern using antidepressants in bipolar disorder, noting that, with the exception of fluoxetine added to olanzapine or lithium, no RCT in bipolar depression has ever shown greater efficacy with an antidepressant plus mood stabilizer versus a mood stabilizer alone. Sidor and MacQueen3 identified a disappointing number-needed-to-treat of 29 when using antidepressants for bipolar depression. Although such observations have spurred interest in various novel molecules other than monoaminergic drugs, they have also effectively halted studies of newer monoaminergic agents that affect novel serotonin receptors or catecholaminergic receptor target combinations.
The field has all but ceased to pursue randomized placebo-controlled trials of any monoaminergic antidepressants developed after 1999 (including escitalopram, desvenlafaxine, duloxetine, levomilnacipran, transdermal selegiline, vortioxetine, and vilazodone) and a handful of those from the mid-1990s (eg, mirtazapine, citalopram, fluvoxamine) or earlier (eg, nortriptyline, phenelzine, isocarboxazid) for which no controlled data exist in bipolar depression. An April 2019 search of the www.clinicaltrials.gov database using the terms “antidepressant” and “bipolar disorder” for active and/or enrolling RCTs returned 42 responses, only two of which involved a previously unstudied monoaminergic antidepressant (vortioxetine [ ClinicalTrials.gov identifier NCT03598868]; and escitalopram [ ClinicalTrials.gov identifier NCT00958633]).
Are more studies with newer monoaminergic antidepressants needed in bipolar depression, or is there already sufficient evidence? Table 1 summarizes existing RCTs.
Evidence Base of Specific Antidepressant Randomized Controlled Trials in Bipolar Disorder
Enrichment: Friend or Foe?
“Enrichment” means selecting patients based on certain baseline characteristics that increase their chances for responding to a given treatment. It is a double-edged sword that trades off generalizability (results do not apply to all-comers) for “cherry-picking” patients most likely to benefit from a particular intervention. In bipoloar II depression, several relapse prevention studies using antidepressant monotherapies have shown significantly better outcomes for selective serotonin reuptake inhibitors (SSRIs) (sertraline, fluoxetine) or venlafaxine than placebo with the important proviso that initial response is robust, as summarized in Table 1. Notably, Altshuler et al.19 found that initial full responders to venlafaxine, sertraline, or bupropion were 6 times more likely than partial responders to remain depression-free or mania-free when continuing antidepressants for 1 year. However, full acute response rates during open-label initial treatment were modest (35% after a first antidepressant trial, 13% with a second, and 9% with third attempt), at a rate of decrement not much different than reported in iterative pharmacotherapy trials of major depressive disorder.
Magnitude of initial acute improvement with an antidepressant appears to heavily influence future course during continued antidepressant therapy. Said differently, partial or incomplete remitters to an acute antidepressant trial are unlikely to improve further with more protracted duration of antidepressant use,19 in contrast to chronic or persistent unipolar depression, where lengthier antidepressant trials often convert partial response to full remissions.
Tricyclic antidepressants20 and venlafaxine16 have been suggested to pose higher inherent risks than SSRIs or bupropion for polarity switch, with speculation that potent noradrenergic effects may be riskier than mainly serotonergic effects, although the putative catecholaminergic effects of bupropion (as well as norquetiapine) seemingly challenge that general hypothesis (Figure 1).
A decision tree for antidepressant use in bipolar disorder.
A number of patient-specific characteristics bear on antidepressant candidacy in bipolar disorder (Figure 1). These are elaborated in the following text.
Evidence-based medicine includes historical outcomes for an individual patient in tailoring their treatment, meaning that a past personal history of antidepressant response carries inarguable weight (antidepressant candidacy is already established). Similarly, a personal history of poor response or polarity switch with antidepressants diminishes candidacy.
Bipolar Subtypes: I Versus II
Antidepressants appear substantially less likely to induce polarity switches in bipolar II than bipolar I depression.21 In fact, evidence for antidepressant efficacy is far better established in bipolar II than bipolar I depression, making sertraline,15 fluoxetine,8 and venlafaxine17,18 monotherapies among the best-studied pharmacotherapies for bipolar II depression (Table 1). Quetiapine holds the largest monotherapy RCT database for bipolar II depression, where post-hoc analyses from US Food and Drug Administration registration trials comparing 300 mg per day (n = 283), 600 mg per day (n = 289), or placebo (n = 204) yielded medium effect sizes at either dose.22 Cautionary admonitions against using antidepressants without concomitant mood stabilizers pertain to bipolar I depression, because in bipolar II depression RCTs, adjunctive lithium lessens antidepressant efficacy and increases study dropout compared to SSRI monotherapy.15
The presence of even low-grade mania symptoms during bipolar depressive episodes has been associated with an increased risk for worsening of mania symptoms during antidepressant exposure.23 The Diagnostic and Statistical Manual of Mental Disorders, fifth edition's24 broadened criteria for the “mixed features” specifier encourages clinicians to carefully distinguish pure from “mixed” depressive episodes, with implications for the appropriateness or inappropriateness of antidepressants.
Randomized data indicate that in patients with past-year rapid cycling, the continued use of adjunctive antidepressants after an initial response triples the likelihood of subsequent depressive relapse.25 In a previously mentioned favorable bipolar II depression relapse prevention trial of fluoxetine, lithium, or placebo,8 exploratory analyses of the small rapid cycling subgroup (n = 25) found no better efficacy than placebo with either drug.26 By contrast, secondary analyses of rapid cycling subgroups27 from an earlier bipolar II depression RCT17 revealed better acute antidepressant responses with venlafaxine monotherapy (n = 12) than lithium (n= 15).
Substance Use Comorbidity
Psychoactive substance misuse may correlate with antidepressant-associated mania. Altshuler et al.15 found a significantly higher risk for switch events in bipolar patients with prior drug abuse or dependence (particularly psychostimulants) after controlling for covariates. Elsewhere, retrospective studies have linked histories of substance misuse with 5- to 7-fold increased risks for antidepressant-associated mania or hypomania.28,29
Multiple Antidepressant Trials
Numerous antidepressant trials have been associated with poorer antidepressant response in bipolar disorder,30 analogous to findings in the unipolar depression literature.
Antidepressants for Purposes Other Than Depression in Bipolar Disorder
There is remarkably scant literature on the use of SSRIs or serotonin-norepinephrine reuptake inhibitors specifically to treat comorbid anxiety disorders in bipolar patients. In an FDA registration trial of quetiapine for bipolar depression, paroxetine (as an active comparator) failed to improve depressive symptoms but on a secondary analysis did improve concomitant anxiety symptoms better than placebo.13 Possible anxiolytic, anti-obsessional, or other psychotropic effects of SSRIs in bipolar disorder are largely unstudied.
Antidepressants may be safe and effective for an identifiable minority of bipolar patients with depression, although most are understudied. Relative to other treatment options, baseline characteristics and subdiagnostic features can help guide individualized candidacy in ways most consistent with evidence-based parameters.
- Pacchiarotti I, Bond DJ, Baldessarini RJ, et al. The International Society for Bipolar Disorders (ISBD) task force report on antide-pressant use in bipolar disorders. Am J Psychiatry. 2013;170(11):1249–1262. https://doi.org/10.1176/appi.ajp.2013.13020185 PMID: doi:10.1176/appi.ajp.2013.13020185 [CrossRef]24030475
- Fornaro M, Anastasia A, Novello S, et al. Incidence, prevalence and clinical correlates of antidepressant-emergent mania in bi-polar depression: a systematic review and meta-analysis. Bipolar Disord.2018;20(3):195–227. https://doi.org/10.1111/bdi.12612 PMID: doi:10.1111/bdi.12612 [CrossRef]29441650
- Sidor MM, MacQueen GM. Antidepressants for the acute treatment of bipolar depression: a systematic review and meta-analysis. J Clin Psychiatry. 2011;72(2):156–167. https://doi.org/10.4088/JCP.09r05385gre PMID: doi:10.4088/JCP.09r05385gre [CrossRef]
- Sachs GS, Nierenberg AA, Calabrese JR, et al. Effectiveness of adjunctive antidepressant treatment for bipolar depression. N Engl J Med. 2007;356(17):1711–1722. https://doi.org/10.1056/NEJMoa064135 PMID: doi:10.1056/NEJMoa064135 [CrossRef]17392295
- Tohen M, Frank E, Bowden CL, et al. The International Society for Bipolar Disorders (ISBD) Task Force report on the nomenclature of course and outcome in bipolar disorders. Bipolar Disord. 2009;11(5):453–473. https://doi.org/10.1111/j.1399-5618.2009.00726.x PMID: doi:10.1111/j.1399-5618.2009.00726.x [CrossRef]19624385
- Tondo L, Vázquez G, Baldessarini RJ. Mania associated with antidepressant treatment: comprehensive meta-analytic review. Acta Psychiatr Scand. 2010;121(6):404–414. https://doi.org/10.1111/j.1600-0447.2009.01514.x PMID: doi:10.1111/j.1600-0447.2009.01514.x [CrossRef]
- Sachs GS, Lafer B, Stoll AL, et al. A double-blind trial of bupropion versus desipramine for bipolar depression. J Clin Psychiatry. 1994;55(9):391–393. PMID:7929019
- Amsterdam JD, Shults J. Efficacy and safety of long-term fluoxetine versus lithium monotherapy of bipolar II disorder: a randomized, double-blind, placebo-substitution study. Am J Psychiatry. 2010;167(7):792–800. https://doi.org/10.1176/appi.ajp.2009.09020284 PMID: doi:10.1176/appi.ajp.2009.09020284 [CrossRef]20360317
- Cohn JB, Collins G, Ashbrook E, Wernicke JF. A comparison of fluoxetine imipramine and placebo in patients with bipolar de-pressive disorder. Int Clin Psychopharmacol. 1989;4(4):313–322. https://doi.org/10.1097/00004850-198910000-00006 PMID: doi:10.1097/00004850-198910000-00006 [CrossRef]2607128
- Prien RF, Kupfer DJ, Mansky PA, et al. Drug therapy in the prevention of recurrences in unipolar and bipolar affective disorders. Report of the NIMH Collaborative Study Group comparing lithium carbonate, imipramine, and a lithium carbonate-imipramine combination. Arch Gen Psychiatry.1984;41(11):1096–1104. https://doi.org/10.1001/archpsyc.1983.01790220086014 PMID: doi:10.1001/archpsyc.1983.01790220086014 [CrossRef]6437366
- Nemeroff CB, Evans DL, Gyulai L, et al. Double-blind, placebo-controlled comparison of imipramine and paroxetine in the treatment of bipolar depression. Am J Psychiatry.2001;158(6):906–912. https://doi.org/10.1176/appi.ajp.158.6.906 PMID: doi:10.1176/appi.ajp.158.6.906 [CrossRef]11384898
- Himmelhoch JM, Thase ME, Mallinger AG, Houck P. Tranylcypromine versus imipramine in anergic bipolar depression. Am J Psychiatry.1991;148(7):910–916. https://doi.org/10.1176/ajp.148.7.910 PMID: doi:10.1176/ajp.148.7.910 [CrossRef]2053632
- McElroy SL, Weisler RH, Chang W, et al. EMBOLDEN II (Trial D1447C00134) Investigators. A double-blind, placebo-controlled study of quetiapine and paroxetine as monotherapy in adults with bipolar depression (EMBOLDEN II). J Clin Psychiatry.2010;71(2):163–174. https://doi.org/10.4088/JCP.08m04942gre PMID: doi:10.4088/JCP.08m04942gre [CrossRef]20122366
- Young LT, Joffe RT, Robb JC, MacQueen GM, Marriott M, Patelis-Siotis I. Double-blind comparison of addition of a second mood stabilizer versus an antidepressant to an initial mood stabilizer for treatment of patients with bipolar depression. Am J Psychiatry.2000;157(1):124–126. https://doi.org/10.1176/ajp.157.1.124 PMID: doi:10.1176/ajp.157.1.124 [CrossRef]10618026
- Altshuler LL, Sugar CA, McElroy SL, et al. Switch rates during acute treatment for bipolar II depression with lithium, sertraline, or the two combined: a randomized double-blind comparison. Am J Psychiatry. 2017;174(3):266–276. https://doi.org/10.1176/appi.ajp.2016.15040558 PMID: doi:10.1176/appi.ajp.2016.15040558 [CrossRef]28135846
- Leverich GS, Altshuler LL, Frye MA, et al. Risk of switch in mood polarity to hypomania or mania in patients with bipolar depression during acute and continuation trials of venlafaxine, sertraline, and bupropion as adjuncts to mood stabilizers. Am J Psychiatry. 2006;163(2):232–239. https://doi.org/10.1176/appi.ajp.163.2.232 PMID: doi:10.1176/appi.ajp.163.2.232 [CrossRef]16449476
- Amsterdam JD, Shults J. Comparison of short-term venlafaxine versus lithium monotherapy for bipolar II major depressive episode: a randomized open-label study. J Clin Psychopharmacol. 2008;28(2):171–181. https://doi.org/10.1097/JCP.0b013e318166c4e6 PMID: doi:10.1097/JCP.0b013e318166c4e6 [CrossRef]18344727
- Amsterdam JD, Lorenzo-Luaces L, Soeller I, Li SQ, Mao JJ, DeRubeis RJ. Short-term venlafaxine v. lithium monotherapy for bipolar type II major depressive episodes: effectiveness and mood conversion rate. Br J Psychiatry. 2016;208(4):359–365. https://doi.org/10.1192/bjp.bp.115.169375 PMID: doi:10.1192/bjp.bp.115.169375 [CrossRef]26892848
- Altshuler LL, Post RM, Hellemann G, et al. Impact of antidepressant continuation after acute positive or partial treatment response for bipolar depression: a blinded, randomized study. J Clin Psychiatry. 2009;70(4):450–457. https://doi.org/10.4088/JCP.08m04191 PMID: doi:10.4088/JCP.08m04191 [CrossRef]19358785
- Peet M. Induction of mania with selective serotonin re-uptake inhibitors and tricyclic antidepressants. Br J Psychiatry. 1994;164(4):549–550. https://doi.org/10.1192/bjp.164.4.549 PMID: doi:10.1192/bjp.164.4.549 [CrossRef]8038948
- Altshuler LL, Suppes T, Black DO, et al. Lower switch rate in depressed patients with bipolar II than bipolar I disorder treated adjunctively with second-generation antidepressants. Am J Psychiatry. 2006;163(2):313–315. https://doi.org/10.1176/appi.ajp.163.2.313 PMID: doi:10.1176/appi.ajp.163.2.313 [CrossRef]16449487
- Young AH, Calabrese JR, Gustafsson U, et al. Quetiapine monotherapy in bipolar II depression: combined data from four large, randomized studies. Int J Bipolar Disord. 2013;1(1):10. https://doi.org/10.1186/2194-7511-1-10 PMID: doi:10.1186/2194-7511-1-10 [CrossRef]25505677
- Frye MA, Helleman G, McElroy SL, et al. Correlates of treatment-emergent mania associated with antidepressant treatment in bipolar depression. Am J Psychiatry.2009;166(2):164–172. https://doi.org/10.1176/appi.ajp.2008.08030322 PMID: doi:10.1176/appi.ajp.2008.08030322 [CrossRef]
- American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Publishing; 2013.
- El-Mallakh RS, Vöhringer PA, Ostacher MM, et al. Antidepressants worsen rapid-cycling course in bipolar depression: a STEP-BD randomized clinical trial. J Affect Disord. 2015;184:318–321. https://doi.org/10.1016/j.jad.2015.04.054 PMID: doi:10.1016/j.jad.2015.04.054 [CrossRef]26142612
- Amsterdam JD, Luo L, Shults J. Efficacy and mood conversion rate during long-term fluoxetine v. lithium monotherapy in rapid- and non-rapid-cycling bipolar II disorder. Br J Psychiatry.2013;202(4):301–306. https://doi.org/10.1192/bjp.bp.111.104711 PMID: doi:10.1192/bjp.bp.111.104711 [CrossRef]
- Amsterdam JD, Wang CH, Shwarz M, Shults J. Venlafaxine versus lithium monotherapy of rapid and non-rapid cycling patients with bipolar II major depressive episode: a randomized, parallel group, open-label trial. J Affect Disord. 2009;112(1–3):219–230. https://doi.org/10.1016/j.jad.2008.03.029 PMID: doi:10.1016/j.jad.2008.03.029 [CrossRef]
- Manwani SG, Pardo TB, Albanese MJ, Zablotsky B, Goodwin FK, Ghaemi SN. Substance use disorder and other predictors of antidepressant-induced mania: a retrospective chart review. J Clin Psychiatry.2006;67(9):1341–1345. https://doi.org/10.4088/JCP.v67n0903 PMID: doi:10.4088/JCP.v67n0903 [CrossRef]17017819
- Goldberg JF, Whiteside JE. The association between substance abuse and antidepressant-induced mania in bipolar disorder: a preliminary study. J Clin Psychiatry. 2002;63(9):791–795. https://doi.org/10.4088/JCP.v63n0907 PMID: doi:10.4088/JCP.v63n0907 [CrossRef]12363119
- Post RM, Leverich GS, Altshuler LL, et al. Relationship of prior antidepressant exposure to long-term prospective outcome in bipolar I disorder outpatients. J Clin Psychiatry. 2012;73(7):924–930. https://doi.org/10.4088/JCP.11m07396 PMID: doi:10.4088/JCP.11m07396 [CrossRef]22480597
Evidence Base of Specific Antidepressant Randomized Controlled Trials in Bipolar Disorder
||Main Findings from Meta-Analyses or Individual RCTs
||Three pooled trials (n= 205) comparing bupropion to other antidepressants: no significant differences in response but lower switch rate (relative risk = 0.34) than with desipramine or venlafaxine (reviewed by Sidor and MacQueen3)
||Comparable desipramine response (5/10) to bupropion (5/9) but higher switch rates (5/10 vs 1/9, respectively); note very small sample sizes7
||From an enriched sample of patients with acute bipolar depression that responded to fluoxetine, patients randomized to fluoxetine (n= 28) vs lithium (n= 26) vs placebo (n= 27) and treated for 1 year, the hazard of relapse was 2.5 times greater with lithium than fluoxetine, with no differences in hypomanic switches8
6-week acute RCT: adjunctive fluoxetine (n= 30, response: 86%) > placebo (n= 29, response: 38%) or adjunctive imipramine (n = 36, response: 57%); concomitant lithium data unavailable9
||6-week acute RCT: imipramine (n= 30, response: 57%) < fluoxetine after 6 weeks (n= 30), same as placebo (n= 29, response:38%); no reported switches9
Similar depression relapse rates but higher mania relapse rates over 30 months with imipramine alone (n= 36) (53%) than lithium alone (n= 42) (26%), or lithium plus imipramine (n= 36) (28%)10
10-week acute RCT: imipramine (n= 39; response: 39%) no different from placebo (n= 43; response: 35%) except when lithium levels subtherapeutic (post-hoc analysis)11
6-week acute RCT (vs tranylcypromine [n= 28]): imipramine (n= 28; response: 48%; 24% switch [lithium cotherapy parameters not specified])12
||10-week acute RCT: paroxetine (n= 35; response: 46%) no different from placebo (n= 43; response: 35%), except when subtherapeutic serum lithium levels (post-hoc analysis)11
8-week RCT: mean change in depression severity with paroxetine (n = 122) no different from placebo (n= 126)13
Significant, comparable improvement either with lithium plus divalproex (n = 16) or either mood stabilizer plus paroxetine (n = 11); less early dropout with paroxetine14
||16-week RCT of bipolar II sertraline monotherapy (n= 45): 73% response, 17% switch; poorer when combined with lithium (n= 46): 48% response, 10% switch15
10-week RCT (vs other antidepressants) added to mood stabilizer; sertraline (n= 76) response: 55%, switch: 11.8% (no placebo)16
||6-week RCT (vs imipramine [n= 28]); tranylcypromine (n = 28): 81% acute response, 12% switch (lithium co-therapy parameters not specified)12
||10-week RCT (vs other antidepressants) added to mood stabilizer; venlafaxine (n= 86) response: 43%; switch: 5.8% (no placebo)16
Bipolar II monotherapy: 12-week open RCT: venlafaxine (n= 43; response: 58%) > lithium (n= 40; response: 20%), no significant change in mania symptom scores; higher dropout with lithium than venlafaxine17
Bipolar II monotherapy: 12-week RCT: venlafaxine (n= 65, response: 68%) > lithium (n= 64, response: 34%); no increased risk for polarity switch though no placebo comparison arm18